Skip to main content

Training SAR-ATR Models for Reliable Operation in Open-World Environments

Publication ,  Journal Article
Inkawhich, NA; Davis, EK; Inkawhich, MJ; Majumder, UK; Chen, Y
Published in: IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing
January 1, 2021

Training deep learning-based synthetic aperture radar automatic target recognition (SAR-ATR) systems for use in an 'open-world' operating environment has, thus far proven difficult. Most SAR-ATR systems are designed to achieve maximum accuracy for a limited set of classes, yet ignore the implications of encountering novel target classes during deployment. Even worse, the standard deep learning training objectives fundamentally inherit a closed-world assumption, and provide no guidance for how to handle out-of-distribution (OOD) data. In this work, we develop a novel training procedure called adversarial outlier exposure (AdvOE) to codesign the ATR system for accuracy and OOD detection. Our method introduces a large, diverse, and unlabeled auxiliary training dataset containing samples from the OOD set. The AdvOE objective encourages a deep neural network to learn robust features of the in-distribution training data, while also promoting maximum entropy predictions for adversarially perturbed versions of the OOD data. We experiment with the recent SAMPLE dataset, and find our method nearly doubles the OOD detection performance over the baseline in key settings, and excels when using only synthetic training data. As compared to several other advanced ATR training techniques, AdvOE also affords significant improvements in both classification and detection statistics. Finally, we conduct extensive experiments that measure the effect of OOD set granularity on detection rates; discuss the implications of using different detection algorithms; and develop a novel analysis technique to validate our findings and interpret the OOD detection problem from a new perspective.

Duke Scholars

Published In

IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing

DOI

EISSN

2151-1535

ISSN

1939-1404

Publication Date

January 1, 2021

Volume

14

Start / End Page

3954 / 3966

Related Subject Headings

  • 4601 Applied computing
  • 4013 Geomatic engineering
  • 3709 Physical geography and environmental geoscience
  • 0909 Geomatic Engineering
  • 0801 Artificial Intelligence and Image Processing
  • 0406 Physical Geography and Environmental Geoscience
 

Citation

APA
Chicago
ICMJE
MLA
NLM
Inkawhich, N. A., Davis, E. K., Inkawhich, M. J., Majumder, U. K., & Chen, Y. (2021). Training SAR-ATR Models for Reliable Operation in Open-World Environments. IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing, 14, 3954–3966. https://doi.org/10.1109/JSTARS.2021.3068944
Inkawhich, N. A., E. K. Davis, M. J. Inkawhich, U. K. Majumder, and Y. Chen. “Training SAR-ATR Models for Reliable Operation in Open-World Environments.” IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing 14 (January 1, 2021): 3954–66. https://doi.org/10.1109/JSTARS.2021.3068944.
Inkawhich NA, Davis EK, Inkawhich MJ, Majumder UK, Chen Y. Training SAR-ATR Models for Reliable Operation in Open-World Environments. IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing. 2021 Jan 1;14:3954–66.
Inkawhich, N. A., et al. “Training SAR-ATR Models for Reliable Operation in Open-World Environments.” IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing, vol. 14, Jan. 2021, pp. 3954–66. Scopus, doi:10.1109/JSTARS.2021.3068944.
Inkawhich NA, Davis EK, Inkawhich MJ, Majumder UK, Chen Y. Training SAR-ATR Models for Reliable Operation in Open-World Environments. IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing. 2021 Jan 1;14:3954–3966.

Published In

IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing

DOI

EISSN

2151-1535

ISSN

1939-1404

Publication Date

January 1, 2021

Volume

14

Start / End Page

3954 / 3966

Related Subject Headings

  • 4601 Applied computing
  • 4013 Geomatic engineering
  • 3709 Physical geography and environmental geoscience
  • 0909 Geomatic Engineering
  • 0801 Artificial Intelligence and Image Processing
  • 0406 Physical Geography and Environmental Geoscience